Model Predictive Control for Multi-Class Traffic Flows

Abstract

In this paper we first present an extension of the macroscopic traffic flow model METANET to multi-class flows. The resulting multi-class model takes into account the differences between, e.g., fast vehicles (cars) and slow vehicles (trucks) including their possibly different free-flow speeds and critical densities. Next, we show how this model can be used in a model-based predictive control approach for coordinated and integrated traffic flow control. In particular, we use Model Predictive Control (MPC) to coordinate various traffic control measures such as variable speed limits, ramp metering, etc. Using a simple benchmark example from the literature we illustrate that by taking the heterogeneous nature of multi-class traffic flows into account a better performance can be obtained.